【机器视觉学习笔记】VS2015 安装 opencv_contrib并测试

平台：Windows 10 20H2
Visual Studio 2015
opencv_contrib-3.4.12

---

参考文章：
添加OpenCV_contrib库至OpenCV3.1.0中（Windows 64位 环境下编译OpenCV3.1.0和OpenCV_contrib）—— MICHAEL_LIU
opencv_contrib安装笔记 —— weijifen000

        Releases - OpenCV上的OpenCV所包含的库只有基础内容，而人脸识别、matlab调用、RGB加工、深层神经网络等内容则需要安装opencv_contrib。

opencv_contrib的获取

        Github：opencv/opencv_contrib

主要工具

1. Cmake：我下载的安装包是cmake-3.21.2-windows-x86_64.msi。
2. opencv 和 opencv_contrib，版本要一致，我安装的版本是3.4.12。
3. Visual Studio：我所用的版本是2015。

编译 opencv

        使用cmake的时候需使用python 2.7，否则之后貌似会有警告。
        使用Anaconda3建立环境后，进入python 2.7环境。
        在python 2.7环境下打开cmake

        Where is the source code内填入opencv解压目录下的 sources 文件夹路径
        Where to build the binaries内填入输出路径

        点击Configure

        选择所用的Visual Studio版本和系统平台

        确认下方框内无红色警告

        若有警告，打开警告位置的日志文件

        查找cmake_download字符串，其后第一个路径是文件放置位置，第二个网址是下载地址

        缺的文件会由于下载失败变成0kb的文件

        下载后改成相同的名字将其替换

        之后再次点击Configure，确认上下方都没红色警告，且提示Configuring done

        之后点击Generate，出现Generating done

编译 opencv_contrib

        在 search 对话框中输入 OPENCV_EXTRA_MODULES_PATH，找到OPENCV_EXTRA_MODULES_PATH后在value中填入opencv_contrib解压目录下的modules路径，反斜杠\要改成正斜杠/

        在 search 对话框中输入 OPENCV_ENABLE_NONFREE ，在value值中勾选以允许使用已申请专利的算法，之后再次点击Configure。

        出现Configuring done后点击Generate

        等待出现Generating done

Visual Studio 编译

点击 Open Project

        选择 生成->批生成

        勾选ALL_BUILD和INSTALL的Debug和Release配置，之后点击生成

        等待编译完成，最终生成的文件不小：

此后该文件夹也可移动到别的地方

配置新项目的环境

新建要用的项目



配置项目的属性

配置选为所有配置，平台为之前所选的平台

添加包含目录

VC++目录->包含目录

添加输出文件夹下install\include文件夹及其子文件夹的路径

添加库目录

VC++目录->库目录

其路径为输出文件夹下/install/x64/vc14/lib文件夹的路径

点击应用

配置调试环境

调试->环境

添加
PATH=D:\Work\OpenCV\opencv-3.4.12_build\install\x64\vc14\bin;%PATH%
路径为输出文件夹下\install\x64\vc14\bin文件夹的路径

点击应用

添加依赖项

分别配置Debug、Release下的 链接器->输入->附加依赖项

加入lib文件夹下的以".lib"结尾的库文件名
其中以"d.dll"结尾的为Debug时需要的库
不带’d’的则为Release时需要的库

这些文件名可用如下python脚本提取，filePath为lib文件夹的路径

import os
filePath = r'D:\Work\OpenCV\opencv-3.4.12_build\install\x64\vc14\lib'

DebugLibList = []
ReleaseLibList = []

for _ in os.listdir(filePath):
    if _.endswith('d.lib'):
        DebugLibList.append(_)
    elif _.endswith('.lib'):
        ReleaseLibList.append(_)

print("DebugLibList:\n")
for _ in DebugLibList:
    print(_)

print("\nReleaseLibList:\n")
for _ in ReleaseLibList:
    print(_)

在本例中
Debug的库为

opencv_aruco3412d.lib
opencv_bgsegm3412d.lib
opencv_bioinspired3412d.lib
opencv_calib3d3412d.lib
opencv_ccalib3412d.lib
opencv_core3412d.lib
opencv_datasets3412d.lib
opencv_dnn3412d.lib
opencv_dnn_objdetect3412d.lib
opencv_dpm3412d.lib
opencv_face3412d.lib
opencv_features2d3412d.lib
opencv_flann3412d.lib
opencv_fuzzy3412d.lib
opencv_hdf3412d.lib
opencv_hfs3412d.lib
opencv_highgui3412d.lib
opencv_imgcodecs3412d.lib
opencv_imgproc3412d.lib
opencv_img_hash3412d.lib
opencv_line_descriptor3412d.lib
opencv_ml3412d.lib
opencv_objdetect3412d.lib
opencv_optflow3412d.lib
opencv_phase_unwrapping3412d.lib
opencv_photo3412d.lib
opencv_plot3412d.lib
opencv_reg3412d.lib
opencv_rgbd3412d.lib
opencv_saliency3412d.lib
opencv_shape3412d.lib
opencv_stereo3412d.lib
opencv_stitching3412d.lib
opencv_structured_light3412d.lib
opencv_superres3412d.lib
opencv_surface_matching3412d.lib
opencv_text3412d.lib
opencv_tracking3412d.lib
opencv_video3412d.lib
opencv_videoio3412d.lib
opencv_videostab3412d.lib
opencv_xfeatures2d3412d.lib
opencv_ximgproc3412d.lib
opencv_xobjdetect3412d.lib
opencv_xphoto3412d.lib

Release的库为

opencv_aruco3412.lib
opencv_bgsegm3412.lib
opencv_bioinspired3412.lib
opencv_calib3d3412.lib
opencv_ccalib3412.lib
opencv_core3412.lib
opencv_datasets3412.lib
opencv_dnn3412.lib
opencv_dnn_objdetect3412.lib
opencv_dpm3412.lib
opencv_face3412.lib
opencv_features2d3412.lib
opencv_flann3412.lib
opencv_fuzzy3412.lib
opencv_hdf3412.lib
opencv_hfs3412.lib
opencv_highgui3412.lib
opencv_imgcodecs3412.lib
opencv_imgproc3412.lib
opencv_img_hash3412.lib
opencv_line_descriptor3412.lib
opencv_ml3412.lib
opencv_objdetect3412.lib
opencv_optflow3412.lib
opencv_phase_unwrapping3412.lib
opencv_photo3412.lib
opencv_plot3412.lib
opencv_reg3412.lib
opencv_rgbd3412.lib
opencv_saliency3412.lib
opencv_shape3412.lib
opencv_stereo3412.lib
opencv_stitching3412.lib
opencv_structured_light3412.lib
opencv_superres3412.lib
opencv_surface_matching3412.lib
opencv_text3412.lib
opencv_tracking3412.lib
opencv_video3412.lib
opencv_videoio3412.lib
opencv_videostab3412.lib
opencv_xfeatures2d3412.lib
opencv_ximgproc3412.lib
opencv_xobjdetect3412.lib
opencv_xphoto3412.lib

测试

来自Opencv3.1.0+opencv_contrib配置及使用SIFT测试 —— 陈纪建

opencv 3中SIFT匹配是在opencv_contrib库中的，这里我们就用它来做一个简单的测试。

#include 
#include   //头文件
#include 

using namespace cv;  //包含cv命名空间
using namespace std;
using namespace xfeatures2d;

int main()
{
	//Create SIFT class pointer
	Ptr<Feature2D> f2d = SIFT::create();

	//读入图片
	Mat img_1 = imread("D:/Work/OpenCV/Workplace/Robot/Robot/1.jpg");
	Mat img_2 = imread("D:/Work/OpenCV/Workplace/Robot/Robot/2.jpg");

	//Detect the keypoints
	vector<KeyPoint> keypoints_1, keypoints_2;
	f2d->detect(img_1, keypoints_1);
	f2d->detect(img_2, keypoints_2);

	//Calculate descriptors (feature vectors)
	Mat descriptors_1, descriptors_2;
	f2d->compute(img_1, keypoints_1, descriptors_1);
	f2d->compute(img_2, keypoints_2, descriptors_2);

	//Matching descriptor vector using BFMatcher
	BFMatcher matcher;
	vector<DMatch> matches;
	matcher.match(descriptors_1, descriptors_2, matches);

	//绘制匹配出的关键点
	Mat img_matches;
	drawMatches(img_1, keypoints_1, img_2, keypoints_2, matches, img_matches);
	imshow("match图", img_matches);

	//等待任意按键按下
	waitKey(0);
}

测试结果如下：